• Rainbow Crops, a Belgium-based startup using AI and multiplex gene editing to improve yield, drought tolerance and other complex traits in corn and other key crops, has raised a €9.7 million ($11.25 million) seed round.
• The round was led by LIFTT EuroInvest, a venture capital vehicle launched as a partnership between the European Investment Bank and Italian VC firm LIFTT.
• Other participants include existing investors AIF (Agri Investment Fund), PINC (the venture arm of food giant Paulig), and Belgian research institute VIB, plus new investors Maia Ventures and Corteva Catalyst.

Why it matters

For crop breeders, some of the most commercially important traits—higher yield, drought tolerance, heat resilience—are the hardest to improve because they are controlled by networks of interacting genes.

Rainbow Crops—which was spun out of Belgian research institute VIB in 2025—is tackling this problem by using a systems biology approach combining three key elements, cofounder and CEO Giacomo Bastianelli tells AgFunderNews.

1 – AI to predict which combinations of genes influence complex crop traits such as drought tolerance or improved yield

2 – Multiplex gene editing to create plants with those combinations coupled with breeding to generate populations of genome edited plants for screening

3 – High-throughput screening to measure how edited plants perform, combined with omics data showing which genes and pathways are active

This genotypic and phenotypic data is then fed back into the model so that it can continuously improve, says Bastianelli, who was brought in to commercialize the tech, which was first developed at VIB by Dirk Inzé, PhD and Hilde Nelissen, PhD.

 “Rainbow Crops is a compelling example of breakthrough innovation in agritech: a category-defining platform that integrates AI, genome editing, and real-world validation to address critical crop breeding challenges.” Edoardo Bianchi, project manager, LIFTT

Using AI to map the genetic network

By using a systems biology approach, Rainbow Crops is trying to understand more holistically what is contributing to a trait, says Bastianelli.

“In general, the traits that we are after are not monogenic [controlled by a single gene]. Viral resistance is likely to be monogenic, or few genes that can be [tweaked]. But yield, biomass or drought tolerance are polygenic traits where you need to modify multiple genes at the same time.

“So we collect data… phenotypic, genotypic, but also omics, transcriptomic data… to get an understanding of the contribution of each edit and map everything out from a systems biology view, how these things all influence each other. Then you can optimize and select the edits that are contributing to the trait.

“It’s this closed loop approach that attracted investors and the Gates Foundation to give us a grant three months ago.”

Rainbow Crops builds a “massive graph” connecting genes to other genes, and genes to traits, using public datasets, scientific literature, and its own generated data. The goal is not just to identify genes associated with a trait, but to understand whether turning a gene up, down, or off is likely to have a positive or negative effect, explains Bastianelli.

It is also moving beyond simple knockouts toward editing regulatory elements to tune gene expression more precisely.

“In the past the modifications were mainly knockouts [of individual genes]. Now we can move into modifying regulatory elements of a gene, so you can tune up or tune down gene expression, for example. If you knock out a gene, you can sometimes [end up with] a detrimental phenotype [an unwanted trait], but if you start playing with gene expression, then you have more control.”

He adds: “You need to understand how gene expression in a tissue is regulated. What are the transcription factors that activate that gene? One human brain or even 100 human brains will not capture those patterns, but AI is particularly well suited to do it.”

‘Rationally designed genetic diversity’

The initial technology developed at VIB was “about combining mainly knockouts with breeding,” says Bastianelli. “It’s not just about gene editing, it’s how we can combine gene editing and breeding to create this population of plants with rationally designed genetic diversity.”

While very large seed companies have built expertise in gene editing, meanwhile, Rainbow Crops probably has an edge in multiplexing, he claims: “We’re talking about not five genes at a time, but 50-100,” including “how you build the vectors” to target such a large set of edits.

Testing and validation

Rainbow Crops has validated its platform through greenhouse trials and a small pilot field trial in corn focused on increased biomass and effects on yield, although the results are still being analyzed. It is also working on sorghum and rice through a Gates Foundation-backed project focused on heat and drought tolerance.

The firm also has access to VIB facilities in Ghent including crop transformation capabilities (getting gene-editing tools into plant cells and then regenerating those edited cells into whole plants) and an automated phenotyping facility that can screen up to 16,000 plants at once using conveyor belts and imaging systems.

The business model

The business model is to partner with seed and breeding companies that have strong germplasm but lack advanced gene-editing and trait-engineering capabilities, says Bastianelli. Rainbow Crops licenses the output of its platform, not the tech itself, earning money from royalties on commercialized varieties.  To generate revenue before varieties reach the market, it also charges upfront R&D fees and milestone payments.  

The company expects to announce a corn collaboration with a seed company soon and is discussing additional partnerships in other crops. Partners would generally bring elite or “semi-elite” germplasm, field-trial capabilities, and regulatory expertise.

Like other plant gene-editing approaches, Rainbow Crops’ platform depends on being able to transform plant cells and regenerate them into whole plants. But Bastianelli argues its multiplex approach reduces the burden by generating many edit combinations from fewer transformation events.

“Even if it doesn’t transform very well, and the efficiency is low, the beauty of this multiplex approach is that with one plant that is transformed, we can generate a population of plants with different combination of edits, and this solves the problem of low transformability, low efficiency of transformation. In general, there are some crops that are difficult to transform, or they regenerate slowly, or not well, so if you have to do 100 transformations to get your population, that will be a pain.”

The company is also exploring whether edited populations could eventually be tested directly in the field rather than first in greenhouses, although this would add regulatory complexity, he says.

The regulatory pathway

As for the regulatory pathways for gene edited crops, they vary by jurisdiction, says Bastianelli, who notes that while there are limits on the number of edits permitted at once in some jurisdictions, this isn’t necessarily something that would thwart players in multiplex editing.

Under the proposed EU framework for plants produced through new genomic techniques (NGT), for example, plants with no more than 20 specified genetic modifications may qualify as NGT-1, the lighter-regulated category treated as equivalent to conventional plants. Plants with more complex modifications would fall into the NGT-2 category and remain subject to existing GMO rules.

“If you develop one plant with four edits and another one with 19 and you cross them, however, that’s OK,” he notes.

Further reading:

🎥 Ag’s new toolkit: AI, genomics, and robotics converge at World Agri-Tech

🎥 Digital twins: Heritable Ag combines AI, genomics and environmental data to slash R&D timelines

Tropic bags $105m to scale gene-edited bananas, deploy TR4 resistant bananas in 2027